Repeatability of Reference Signal Received Power Measurements in LTE Networks. Hayder A Abdulrasool Khzaali

Transcription

1 Repeatability of Reference Signal Received Power Measurements in LTE Networks By Hayder A Abdulrasool Khzaali Bachelor of Science Computer Engineering and Information Technology University of Technology, Baghdad, Iraq, 2008 A thesis submitted to the College of Engineering at Florida Institute of Technology In partial fulfillment of the requirements for the degree of Master of Science In Computer Engineering Melbourne, Florida May 2017 i

2 Copyright 2017 Hayder Khzaali All Rights Reserved The author grants permission to make single copies

3 We the undersigned committee hereby approve the attached thesis, Repeatability of Reference Signal Received Power Measurements in LTE Networks By Hayder A Abdulrasool Khzaali Josko Zec, Ph.D. Associate Professor Electrical and Computer Engineering Ivica Kostanic, Ph.D. Associate Professor Electrical and Computer Engineering Ersoy Subasi, Ph.D. Assistant Professor Engineering Systems Samuel P. Kozaitis, Ph.D Professor and Department Head Electrical and Computer Engineering

4 Abstract Title: Repeatability of Reference Signal Received Power Measurements in LTE Networks Author: Hayder Khzaali Major Advisor: Josko Zec, Ph.D. Reference Signal Received Power (RSRP) reports are routinely used to monitor and benchmark the coverage in LTE cellular networks of fourth generation. Such measurements also trigger idle and active mode mobility decisions between two LTE cells or two inter-technology cells. Therefore, these measurements must be accurate and repeatable to facilitate both proper network functioning and network optimization. This thesis presents a repeatability study on raw RSRP measurements collected in the LTE 700MHZ frequency band in a commercial LTE network. This process is facilitated through the common RF drive-test method using a scanner-based tool as the measuring device. After gathering all the required data with the scanner, the main measurement statistical metrics are calculated in the analysis phase. The repeatability of the RSRP measurements is examined comparing statistics from data collected over identical drive route across different days. Based on the limited collected data set, good temporal repeatability and stability of the LTE RSRP measurements is confirmed allowing optimization to be based on a single LTE drive test. iii

5 Table of Contents Table of Contents... iv List of Figures... vi List of Tables... vii Acknowledgement... viii Dedication... ix Chapter One Introduction and Literature Review Introduction The Cellular Wireless Evolution Thesis Motivation Third Generation Partnership Project (3GPP) Thesis Outline... 6 Chapter Two Introduction to LTE System Introduction LTE Network Architecture Evolved packet Core (EPC) Evolved Universal Terrestrial Radio Access Network (E-UTRAN) User Equipment (UE) User Equipment Categories User Equipment Mode of Operation LTE Frequency bands and Spectrum Allocations TDD and FDD LTE Frequency Bands LTE Bands Overview LTE Channel Bandwidth Range Chapter Three LTE Radio Interface Introduction LTE Modulation Schemes Multiple Access Techniques TDD and FDD Modes Operation of TDD and FDD Modes Orthogonal Frequency Division Multiplexing (OFDM) Orthogonal Frequency Division Multiple Access Single Carrier Frequency Division Multiple Access LTE Air Interface protocol Stack iv

6 3.6 LTE Channel Structure Logical Channels Transport Channels Physical Channels LTE Measurements Chapter Four Measurement Procedure Introduction Measurement Tools Drive Test Procedure Measurements Analysis Data Binning Averaged RSRP Analysis Comparison of RSRP Measurements Results for Comparison of RSRP Measurements RSRP difference distribution Results from RSRP difference distribution Chapter Five Conclusion and Future Work Conclusion Further Investigation Chapter Six References...77 Chapter Seven Appendices...81 Appendix A Abbreviation Appendix B MATLAB Code used for Analysis Appendix C Utilized Applications v

7 List of Figures FIGURE 1. CELLULAR WIRELESS EVOLUTION [1]...4 FIGURE 2. DATA AND VOICE TRAFFIC MEASUREMENTS FOR GLOBAL MOBILE TELECOMMUNICATIONS NETWORK FROM ERICSON THROUGH THE PERIOD FROM JAN 2007 TO JULY 2011 [4]...5 FIGURE 3. LTE ARCHITECTURE [7]...8 FIGURE 4. PRIMARY COMPONENTS OF EVOLVED PACKET CORE [3] FIGURE 5. E-UTRAN ARCHITECTURE [3] FIGURE 6. ENB MAIN FUNCTIONS AND CONNECTIONS TO OTHER LOGICAL NODES [10] FIGURE 7. USER EQUIPMENT INTERNAL ARCHITECTURE [3] FIGURE 8. FREQUENCY BAND DEFINITION FIGURE 9. LTE CHANNEL BANDWIDTH OPTIONS [18] FIGURE 10. LTE MODULATION SCHEMES [3] FIGURE 11. MULTIPLE ACCESS TECHNIQUES [2] FIGURE 12. FDD AND TDD OPERATION MODES [3] FIGURE 13. INTER SYMBOL INTERFERENCES REDUCTION BY USAGE OF SUB-CARRIERS [3] FIGURE 14. FREQUENCY AND TIME IMPLEMENTATION IN OFDMA [3] FIGURE 15. OFDMA WAVEFORM EXAMPLE. (A) INDIVIDUAL SUB-CARRIERS AMPLITUDES. (B) RESULTING OFDMA WAVEFORM AMPLITUDE. (C) RESULTING OFDMA WAVEFORM POWER FIGURE 16. OFDMA TRANSMITTER AND RECEIVER [20] FIGURE 17. SC-FDMA TRANSMITTER AND RECEIVER WITH FREQUENCY DOMAIN SIGNAL GENERATION [20] FIGURE 18. OFDMA VS SC-FDMA ACCESS MODES [22] FIGURE 19. LTE AIR INTERFACE PROTOCOL STACK [3] FIGURE 20. FRONT VIEW OF SEEGULL EX+ RECEIVER [27] FIGURE 21. ROUTE MAP FIGURE 22. BINNING IDEA [26] FIGURE 23. AVERAGED RSRP FOR A SUBSET OF GEOGRAPHICAL BINS (ALL BINS) FIGURE 24. AVERAGED RSRP FOR A SUBSET OF GEOGRAPHICAL BINS (ONLY BINS WITH MORE THAN 5 READINGS) FIGURE 25. PAIR-WISE RSRP DIFFERENCES FOR A SUBSET OF ALL BINS FIGURE 26. PAIR-WISE RSRP DIFFERENCES FOR A SUBSET OF BINS WITH AT LEAST 5 RSRP MEASUREMENTS FIGURE 27. ANOVA GRAPH FOR A SUBSET OF ALL BINS FIGURE 28. ANOVA GRAPH FOR A SUBSET OF BINS WITH AT LEAST 5 RSRP MEASUREMENTS FIGURE 29. F-DISTRIBUTION CURVE FIGURE 30. TYPICAL PROBABILITY DENSITY OF THE RSRP DIFFERENCES FOR ALL BINS AND BINS WITH MORE THAN 4 READINGS FIGURE 31. DISTRIBUTION OF STANDARD DEVIATIONS OF RSRP DIFFERENCES FOR ALL BINS FIGURE 32. DISTRIBUTION OF STANDARD DEVIATIONS OF RSRP DIFFERENCES FOR BINS WITH 5+ RSRP MEASUREMENTS vi

8 List of Tables TABLE 1. USER EQUIPMENT CATEGORIES [12] TABLE 2. LTE OPERATING BANDS TABLE 3. CHANNEL QUALITY INDICATOR [21] TABLE 4. RSRP MEASUREMENT REPORTS MAPPING [23] TABLE 5. RSRQ MEASUREMENT REPORTS MAPPING [23] TABLE 6. ANOVA TABLE FOR A SUBSET OF ALL BINS TABLE 7. ANOVA TABLE FOR A SUBSET OF BINS WITH AT LEAST 5 RSRP MEASUREMENTS TABLE 8. F DISTRIBUTION TABLE FOR 0.05 [30] TABLE 9. SUMMARIZED RESULTS vii

9 Acknowledgement I would like to seize this opportunity and express my sincere regards and appreciation to all those who provided me with the tools and knowledge to complete this work. First, I would like to thank my advisor Dr. Josko Zec whose office door was always open whenever I ran into a trouble spot or had any question in mind about my thesis. He provided me with the guidance, and knowledge, and steered me in the right direction whenever he thought I needed it, and it has been an honor to be under his supervision. Second, a special gratitude I would gave to Dr. Ivica Kostanic for his valuable contribution and support all over this journey. I would also like to express my regards to Mr. Dale Bass, and Mr. Loc Acarya from PCTEL Inc, for their assistance and help. Finally, I must declare my very profound gratitude to my parents and my wife for their unfailing support and continuous encouragement throughout my years of study and through the process of researching and writing this thesis. This achievement would not have been possible without them. Thank you all viii

10 Dedication To my mother, Mrs. Hadiya Al-Daffaie, the love of my life. To my Father, Mr. Adnan Khzaali, the idol of my life. To my Brother, Mr. Ahmed Khzaali, the source of my strength. To my lovely wife, Sarah Alyousefi, and my beautiful daughters Saba and Lara, my everything in life. To my lovely sisters, Mrs. Enas khzaali, Mrs, Zena Khzaali, and Mrs. Zahraa Khzaali, the source of my inspiration To all my relatives and friends Finally, to my home country Iraq, and my employer for granting me this opportunity to complete my higher education in this great country. ix

11 Chapter One Introduction and Literature Review

12 1.1 Introduction The growing public demand for Internet services through mobile telecommunication devices was not a surprise to many big cellular companies, especially with the amount of data traffic the social media, online gaming, and video conferencing applications consume every second around the world. With the advancement of LTE networks and MIMO technologies, it becomes possible to meet most of these demands that were impossible to achieve with previous generation networks like GSM (Global System for Mobile Communication) and WCDMA. However, a good LTE cellular coverage is needed to make this happen, and these companies have to face many challenges to accomplish this task. The most important parameter that determines the cellular signal power and quality in a LTE network is RSRP; therefore, one of the essential measurements performed in any LTE cellular telecommunication system is measuring this Key Performance Indicator (KPI). In order for any radio communication device to work properly, its radio circuits have to operate within certain limits defined by the design standards. In other words, it can only operate with a specific range of received signal levels. Both the base stations (cellular towers) and the mobile radio devices (usually cell phones) have to send enough power into the wireless medium to establish and maintain an acceptable voice quality without polluting the wireless frequency channel with excessive power that will affect other entities of communication in the same vicinity. Receivers, on the other hand, have to have a fairly sophisticated signal modulation schemes to be able to retrieve the original signal form, even if the received RSRP is fairly low. The mobile communication industry is a very competitive business, and its success depends on the Quality of Service (QoS) provided. QoS elements can be divided into network accessibility, voice quality, dropped calls, and coverage span. All the above factors rely heavily on the RSRP level. Therefore, measuring RSRP readings is an essential task in any highly dynamic cellular communication environment. It is also necessary to make sure that these measurements are, on average, repeatable to provide robust and efficient performance that meets the subscriber s needs. Power measurements had been a challenge in the past due to the lack of novel techniques, but with today s equipment, it is fairly easy to obtain these measurements 2

13 with common Radio Frequency (RF) engineers tools. These instruments are adapted to be used in the field and are called drive test systems. RF engineers use this method in every day practice to examine the performance of cellular networks. Many network parameters are being reported in the process, and the aim here is to provide those values from the user point of view. After this step, the post-processing phase will be conducted on the collected data for optimization purposes, if necessary. 1.2 The Cellular Wireless Evolution Since the 1G mobile system, which was implemented in 1981, and the cellular wireless networks have witnessed huge development processes, and Figure 1 [1] shows the cellular wireless evolution. Cell phones, on the other hand, have undergone major development phases that have led to evolutional changes. The large cell phones that used the analogue technology were replaced in the early 1990s by the Second Generation (2G) digital system cell phones, due to the need for convenient mobile devices that could fit into every person s pocket. Additionally, these digital cell phones utilized the wireless spectrum more efficiently. New applications emerged that required data transfer through the cellular network and with people becoming more mobile, the Third Generation (3G) took over, providing significant improvements over the second generation in the overall capability of the system. With more and more social media applications being deployed, in addition to the online gaming and video streaming applications, providing sufficient bandwidth to the end user became a priority for all major wireless cellular network providers. Therefore, the Fourth Generation (4G) was introduced in 2010, and is the system used today. The LTE system or the 4G provides much higher throughput and further spectrum efficiency utilization in comparison to previous generations. Later on, Third Generation Partnership Project (3GPP) developed an upgrade for the LTE system that further enhances its capability, and it is called the LTE Advanced [2]. Theoretically, LTE Advanced should provide peak data rates of 1000 Mbps and 500Mbps for the downlink and uplink, respectively [3]. 3

14 In the 2020s, 3GPP is hoping to start implementing the 5th Generation (5G), which is already under developement, that will take the data throughput and capacity to a whole new level and the Internet of Things (IoT) era will soon come to reality. Figure 1. Cellular Wireless Evolution [1] 1.3 Thesis Motivation Mobile telecommunications network traffic was dominated by voice calls in the early years while mobile data growth was initially slow up to 2010 when LTE was introduced, and then everything began to change rapidly. Figure 2 below illustrates a global view of the total network traffic data and voice measurements in petabyte from January 2007 until July 2011, according to Ericson [4]. It is easy to see from Figure 2 that the data traffic amount built up quickly by a factor of over a hundred, and today this trend seems to continue. Therefore, it is of the utmost importance to monitor the performance of established LTE networks to enhance and optimize its parameters if needed, to satisfy this huge increase in mobile telecommunication devices. 4

15 This thesis goal is to measure the repeatability of RSRP measurements from a leading nationwide cellular telecommunication company in the USA, through a common RF engineers DT method using SEEGULL EX + scanning receiver as a data collecting tool. The recorded data will be analyzed using post-processing statistical applications, specifically Matlab and Excel, to provide efficient and authentic information about the network. There are many applications for RF engineers associated with this information such automatic cell planning (ACP) or propagation model optimization (PMO). Figure 2. Data and Voice Traffic Measurements for Global Mobile Telecommunications Network from Ericson through the Period from Jan 2007 to July 2011 [4] 1.4 Third Generation Partnership Project (3GPP) Seven telecommunications standard development organizations are consolidated to form this association (TTC, ETSI, ARIB, ATIS, CCSA, TSDSI, TTA) and are known as the Organizational Partners; they provide the specifications and reports that designate 5

16 3GPP specifications and technologies like WCDMA, UMTS, LTE and LTE Advanced in what is called Releases [5]. Their goal is to cover all cellular network technologies by providing complete system specifications, including radio access and core access as well as services capabilities, such as security and quality of services. They also provide solutions for internetworking with Wi-Fi networks or for non-radio access to the core network. The 3GPP studies and specifications are contributed from member companies in the working group at the technical specifications level, and they can be further sub divided into three groups: RAN or Radio Access Networks SA or Services and systems Aspects CT or Core and network Terminals These technical groups meet on regular bases every three months to present, discuss and request approval for their work, and this meeting is called the quarterly plenary meeting. There are specific responsibilities for each Technical Specification Group (TSG) that can be found in [6]. 1.5 Thesis Outline This thesis is organized as follows: Chapter One is an introduction and a brief cellular communication background. Chapter Two explains the LTE system architecture and its main nodes along with LTE bands overview. Chapter Three describes the air interface between the User Equipment (UE) and the enodeb, explaining modulation schemes, LTE protocol stacks and channel structure. Chapter Four provides details about the methodology and feasible work done in this thesis. Chapter Five includes the conclusion and suggestions for future work. The last two, chapters Six and Seven, will list the references and appendices respectively. 6

17 Chapter Two Introduction to LTE System 7

18 2.1 Introduction An overview of the LTE system architecture and the functionality of its main nodes are presented in this chapter. User equipment (UE) categories and their capabilities are also explained. Finally, the interfaces and protocols between the system components are discussed as well. 2.2 LTE Network Architecture The LTE network architecture is composed of three main components: Evolved Packet Core (EPC) Evolved Universal Terrestrial Radio Access Network (E-UTRAN) User Equipment (UE) A detailed representation of the LTE network architecture showing multiple enodebs and a clearer vision of the core network is illustrated in Figure 3 [7]. Figure 3. LTE Architecture [7] 8

19 All the main elements comprising the LTE network will be briefly explained. It is important to notice here that the enodebs have many of the tasks that in previous generations were associated with the Node B/BSC and BTS/RNC. While BTS represents the base station of the second generation wireless technologies like GSM and CDMA, Node B is the 3rd Generation counterpart such as Worldwide Interoperability for Microwave Access (WiMAX) and UMTS. The same is true for the Base Station Controller (BSC) and Radio Network Controller (RNC) which are associated with the 2nd and 3rd generations, respectively. One of the new functions of enodbs is that they can now be connected to each other via the X2 link, which in turn, will help with handovers and other traffic management functions. Several enodebs can be connected to one mobility management entity (MME) or serving gateway (SGW), and the number of the last two in any LTE network can vary depending on the network configuration. There can also be a couple of pending Gateways (PGW) in the LTE network, depending on the number of assignments associated with each PGW Evolved packet Core (EPC) The Core Network or the EPC communicates with packet data networks from the Internet or the Intranet through corporate networks or the IP multimedia subsystem. Figure 4 shows the main components of the Evolved Packet Core [3]. 9

20 Figure 4. Primary components of Evolved Packet Core [3] A brief description of each component in the architecture above is mentioned below: The Home Subscriber Server (HSS): It is the central data base that has the entire network operator s subscriber s information, and this unit has the same role in both GSM and UMTS. The Packet Data Network (PDN) Gateway (P-GW): It connects to the Internet through serving gateway interface (SGI), and every packet data network is identified by an access point name (APN). The same entity in GSM is called the Serving GPRS Support Node (SGSN), and in UMTS is named the Gateway GPRS Support Node (GGSN). The serving gateway (S-GW): The role of this unit is basically to behave like a router that forwards the data between the PDN gateway and the base station. The mobility management entity (MME): This unit is the brain of the EPC, as it resides in its control plane, and is responsible for the session status management, paging, authentication, mobility in 3GPP, 2G and 3G nodes, Barrier management functions, roaming, and others that can be found in[8] [9]. 10

21 The Policy Control and Charging Rules Function (PCRF): Although it is not shown in the figure above because it is not technically part of the EPC, it is responsible for the basic policy control decision making for the entire LTE network. It is also in charge of the flow-based charging operations in the Policy Control Enforcement Function (PCEF) that is located in the PGW. It can also be noticed that there are two interfaces between the serving gateway and the PDN gateway, namely S5/S8, because they have two different implementations. S8 interface is used when the two devices are in different networks, and S5 is utilized when they are in the same network Evolved Universal Terrestrial Radio Access Network (E-UTRAN) The architecture of the E-UTRAN is shown in Figure 5 [3]. The E-UTRAN is in charge of handling all the wireless communications between the UE and the EPC, and it has only one component which is the enhanced or evolved NodeB (enb). Figure 5. E-UTRAN Architecture [3] 11

22 Every enb is considered as a base station that controls mobile devices in one or more cells, and the base station that is connected to a UE is called the serving enb. The three main elements that define the enb are the antenna system, radio components, and digital processing units. Therefore, enb provides the physical radio connection between the UE and the Cellular tower. However, it does more than that; it provides all radio resource management functions such as radio mobility, admission, and bearer control. The enb also runs the scheduler, making it responsible for managing dynamic resource allocation to its UEs in both downlink and uplink in the LTE network. The enb transmits radio commands to all its UEs via the downlink and receives back their transmissions on the uplink. Figure 5 shows that that the enb is connected to the EPC through the S1 signaling interface and they are connected to each other via the X2 interface. The enb helps in sending all user plane data to the appropriate SGW, and it makes sure that data are encrypted while being sent over the radio interface. It also does header compression by compressing IP packet headers for more efficient use of the radio spectrum. Moreover, one of the most important functions of the enb is to gather, manage, and analyze the UEs radio signal level readings during mobility inside the network. Based on these measurements, a decision will be made by the current enb whether the UE should be handed over to other enbs in the vicinity. Summarized functions of the enb showing connections to other logical nodes are presented in Figure 6 [10]. Two sub interfaces can be derived from the S1-interface, which are the S1-MME and the S1-U. While the first carries the control information or control plane protocols, the other one is responsible for routing the traffic data. The X2 link is utilized for handover and data transfer coordination. 12

23 Figure 6. enb Main Functions and Connections to other Logical Nodes [10] The modulation schemes used by the enbs are the Orthogonal Frequency-Division Multiple Access (OFDMA) and Single-Carrier Frequency-Division Multiple Access (SC- FDMA) for downlink and uplink respectively, which will be explained later User Equipment (UE) The user equipment is the actual communication device, and it has the same architecture in GSM and UMTS as shown in Figure 7 [3]. It can be divided into Terminal Equipment (TE), that terminates data streams, and Mobile Termination (MT), which takes care of communication tasks. The MT can be a plug-in LTE card inside a laptop, which will be considered as the TE in this case. The Subscriber Identity Module card, generally known as the SIM card, and originally known as the Universal Integrated Circuit Card (UICC), is a smart card. This card executes an application called the Universal SIM (USIM) [11] that stores specific 13

24 user data, such as home network identity and phone number, and also uses its stored security keys to handle security related tasks. The LTE supports all devices that use USIM releases 99 and beyond, but it does not support earlier releases of GSM that used the Subscriber Identity Module (SIM). Figure 7. User Equipment Internal Architecture [3] LTE also supports mobile devices that operate with the new version of the Internet Protocol (IPV6), as well as the old version IPV4, or even those which use both stacks at the same time. Alternatively, the mobile devices can also receive both IP addresses, each coming from different packet data networks that it communicates with, like Internet and Intranet, if the mobile device and the network support both protocol versions User Equipment Categories For effective communication between the enodeb and the mobile terminal, 3GPP has created different UE categories that define the limitations and abilities of the UEs, including maximum downlink and uplink speed. Table1 [12] below illustrates UE categories 1 to 10 with their 3GPP release, number of supported MIMO antennas, and maximum downlink and uplink speed, along with some UE examples that support the specific category. 14

25 UE Category 3GPP Release Table 1. User Equipment Categories [12] Max Downlink speed Max Uplink speed Category 1 Release Mbit/s 5.2 Mbit/s Category 2 Release Mbit/s 25.5 Mbit/s Category 3 Release Mbit/s 51.0 Mbit/s Category 4 Release Mbit/s 51.0 Mbit/s Category 5 Release Mbit/s 75.4 Mbit/s Category 6 Release 10 (LTE- Advanced) Category 7 Release 10 (LTE- Advanced) Release 10 Category 8 (LTE- Advanced) Release 10 Category 9 (LTE- Advanced) Release 10 Category 10 (LTE- Advanced) Mbit/s 51.0 Mbit/s Mbit/s Mbit/s No. of MIMO 1 N/A 2 N/A Mbit/s Mbit/s 8 N/A Mbit/s 51.0 Mbit/s Mbit/s Mbit/s Examples of supported devices 2 Original Moto X, iphone 5 2 Nexus 5, Moto G 4G, Moto X (2nd Generation), iphone 6, MI Note 4G, Samsung Galaxy S5, LG G3 4 N/A 2 or 4 Huawei Honor 6, Samsung Galaxy S6 LTE-A, LG G3 LTE- A, Nexus 6, Axon 7 2 or 4 N/A 2 or 4 Galaxy Note 7, Galaxy Note 5, iphone 7 2 or 4 Galaxy Note 7 From Table 1, one can notice that Category 4,6, and 9 mobile devices are the most common, while devices that operate within Category 3 are gradually becoming outdated. Categories 6 and beyond fall within the LTE advanced releases that support the Carrier aggregation special characteristic. This techinque is used for better spectrum utilization by enabiling one or more Component Carriers to be aggregated with the Primary Component Carrier (PCC) to increase the bandwidth, and thereby, the bitrate. The UE can aggregate maximum up to 5 carriers, 1 PCC and 4 Secondary Component carriers (SCC). However, it is not possible to configure a UE with more Uplink component 15

26 carriers than Downlink component carrirs, but the reverse is possible. UE Categories 6 and 7 will indicate Category 4 as well, UE Category 8 will also idicate Category 5, and UE Category 9 will indicate Category 6, and 4 and UE Category 10 will indicate Category 7, and 4 [13] User Equipment Mode of Operation There are four modes of operation that the UE can operate with when it is connected to the Evolved Packet System (EPS), which are defined below [14] where PS and CS are abbreviations for Packet Switching and Circuit Switching respectively: PS mode 1: The user equipment only registers to EPS services and the UE usage setting is voice centric. PS mode 2 : The user equipment only registers to EPS services but the UE usage setting is data centric CS/PS mode 1 : The user equipment is registerd to both EPS and non-eps services but the UE usage voice centric of the non-eps services is preferred. CS/PS mode 2 : The user equipment is registerd to both EPS and non-eps services but the UE usage data centric of the EPS services is preferred. 2.3 LTE Frequency bands and Spectrum Allocations The LTE network supports a wide range of frequency bands due to the growing number of cellular devices around the world. There is a wide range of frequency bands used for LTE Time Devision Dublexing (TDD) and LTE Frequency Devision Dublixing (FDD), which will be explained in Chapter 3. Most of the current LTE frequency bands are already in use in many cellular systems while other new bands are being introduced in regions where the current spectrum is being allocated elsewhere. 16

27 2.3.1 TDD and FDD LTE Frequency Bands TDD spectrum requires only a single band for both uplink and downlink, as they are separated in time, while FDD demands paired bands, one for the uplink and the other for the downlink. As a consequence, there are different band allocations for FDD and TDD and in some cases, although unlikely, both FDD and TDD transmissions could exist in a particular frequency band and overlap between the two can happen. Therefore, romaing UEs may encounter both types of transmission on the same band, and the challenge is to be able to determine whether a TDD or FDD transinssion must be made in that particular LTE band in its current location. Figure 8 illustrates the definition of the LTE frequency band [15]. A large portion of the radio spectrum has been reserved for the FDD LTE use. In this scheme, the bands are paired to permit simultaneous transmission on both frequencies. There is sufficient seperation between them in order to avoid partial collisions that would lead to signal distortion and impair the performance of the receiver. On the other hand, the uplink and downlink share the same link in the TDD LTE scheme, and therefore, the bands are unpaired because they are multiplexed in time. Additions to the LTE frequency bands happen as a result of the International Telecommunication Union (ITU) regulatory meetings as they discuss the need for extra spectrum. The driving force behind the ITU meetings comes from the continual growing need for mobile communications.however, the the overall LTE spectrum is also limited and therefore, the new spectrum allocations are relatively small, between 10 to 20 MHz in bandwidth. This will cause another problem when LTE-Advanced is introduced because its bandwidth requirement is 100MHz. Therefore, the channel aggregation should occur over a wide ranage of frequencies, and this is considered a significant technological issue. Table 2 illustrates the LTE operating bands [15][16]. 17

28 LTE Figure 8. Frequency Band Definition Operating Bands Table 2. E-UTRA Operating Band Uplink (UL) in MHz operating band BS receive UE transmit Downlink (DL) in MHz operating band BS transmit UE receive Duplex Mode WIDTH OF BAND (MHZ) DUPLEX SPACING (MHZ) BAND GAP (MHZ) FUL_low FUL_high FDL_low FDL_high FDD FDD FDD FDD FDD FDD FDD FDD FDD FDD FDD FDD FDD FDD

29 Reserved Reserved Reserved Reserved FDD FDD FDD FDD FDD FDD FDD FDD FDD FDD FDD FDD 30/ FDD FDD N/A FDD FDD FDD TDD 20 N/A N/A TDD 15 N/A N/A TDD 60 N/A N/A TDD 60 N/A N/A TDD 20 N/A N/A TDD 50 N/A N/A TDD 40 N/A N/A TDD 100 N/A N/A TDD 194 N/A N/A 19

30 TDD 200 N/A N/A TDD 200 N/A N/A TDD 100 N/A N/A Note: Band 6 is not applicable LTE Bands Overview As the demands for more spectrum increase, more numbers of bands must be allocated to accommodate the pressure on the spectrum. However, these band allocations were not the same around the globe due to the different regulatory provisions in many countries. Therefore, it was not possible to have the same allocations globally, and as result, bands may overlap with each other in some cases which causes roaming in LTE to have some limitations since some UEs lack the ability to access similar frequencies at the same time. Below are some notes that accompany the tabulations of some LTE bands [17]: LTE Band 1: It is a paired band used for FDD that was defined for the 3GPP release 99. LTE Band 4: This band was brought forward by the World Radio Conference WCR-2000, where approval of all international spectrum allocations take place, as the new band for the Americas. The downlink of this band overlaps with one of Band 1 which facilitates roaming. LTE Band 9: This band was introduced to be only used in Japan. It partially overlaps with Band 3 but has different band limits. This again makes roaming easily achievable and many terminals or UEs are designed to operate with dual bands 3+9. LTE Band 10: This band is not available everywhere, and it is an extension to Band 4 as it gives 15 MHz increase, from 40 to 60 MHz (paired). LTE Band 11: Although this band was identified by 3GGP to be used in Japan this 1500 Mhz band is globally allocated to the mobile services as well on a co-primary bases. LTE Band 12: As a result of the Digital Dividend, which is a term that refers to a gap between regions that have access to modern 20

31 communications technology and other areas that have little to no access at all, this band was released. It was used for broadcasting service in the past. LTE Band 13: The same cause for Band 12 releasing had led to the release of this band and it was also previously used for broadcasting purposes. In this band, the uplink has higher frequency than the downlink because of the reversed duplex configuration. LTE Band 14: Again this band was used previously for broadcasting and was released because of the Digital Dividend. The downlink has lower frequency than the uplink because of the reversed duplex configuration. LTE Band 15: This band was intended to be used in Europe and was defined by the ETSI (European Telecommunications Standards Institute). In this band two TDD bands are combined to provide one FDD band. This band has been adopted by 3GPP. LTE Band 16: same as Band 15. LTE Band 17: This band has also been released as a result of the Digital Dividend and was previously used for broadcasting. LTE Band 20: The uplink in this band also uses higher frequency than the downlink because of the reversed duplex configuration from the standard. LTE Band 21: This is a 1500 MHz band originally intended for use in Japan and it was identified by 3GPP. LTE Band 24: The duplex configuration for this band is the same as Bands 13, 14 and 20. Therefore, the uplink has a higher frequency than the downlink. LTE Band 33: This band was defined by 3GPP in release 99 for the unpaired spectrum or TDD use. LTE Band 34: This band is also introduced by 3GPP specifications in release 99 for TDD use. LTE Band 38: This band is in the center between the downlink and the uplink pairs of the LTE Band 7. Although 3GPP can introduce bands for LTE use or any other mobile service, the global allocations of these bands can only be decided by ITU. Only after that happens, each country s administration may allocate a portion from the overall spectrum for their own use, in their territory, and 3GPP has no legal basis to do that. Frequency bands can 21

32 be allocated using primary and secondary bases since the primary users always have prior access to the band, and the secondary users, in most cases, use the band provided without causing interference to the primary users. 2.4 LTE Channel Bandwidth Range LTE supports 6 bandwidth options, ranging from 1.4 MHz up to 20 MHz, as shown in Figure 9 [18]. This range provides a great deal of flexibility when choosing the right bandwidth for the operator, depending on its situation [19]. Figure 9. LTE Channel Bandwidth Options [18] 22

33 Chapter Three LTE Radio Interface 23

34 3.1 Introduction In this chapter, LTE modulation techniques are illustrated, including TDD, FDD, OFDMA and SC-FDMA modes will be explained. Next, LTE protocol stack and channel structure with a brief explanation of each main channel type and its sub-categories are presented and finally some important LTE measurement parameters, such as RSRP and RSRQ, are mentioned as well. 3.2 LTE Modulation Schemes Four modulation schemes are used in LTE. The first scheme used for control streams is the Binary Phase Shift Keying (BPSK) which sends one bit at a time, using two amplitudes of -1 and +1, or two phases of 180 and 0 degress. This scheme is not used for regular data transmission. The second scheme is the Quadrature Phase Shift Keying (QPSK) which sends two bits at a time with four different states. The third one is the 16 Quadrature Amplitude Modulation (16-QAM) that sends four bits at a time, and this scheme has 16 different states that can vary in amplitude and phase. The Last scheme is the 64 Quadrature Amplitude Modulation (64-QAM) which transmits 6 bits at a time and has 64 different states. All modulation schemes are illustrated in Figure 10[3]. Figure 10. LTE Modulation Schemes [3] 24

35 All of these modulation techniques mentioned above, except for BPSK which is not used for normal data transmission, are supported by LTE in the downlink and the uplink. However, the 64-QAM modulation scheme is optional on the uplink. On the downlink, the LTE system supports Quadrature Phase Shift Keying (QPSK) and Quadrature Amplitude Modulation for both 16QAM and 64QAM. The same modulation types are supported on the uplink. However, the use of 64QAM will depend on the UE capability, and for some mobiles, support of 64 QAM is optional [20]. The selection of any modulation scheme depends on the Signal to Interferences plus Noise Ratio (SINR) of the measured signal. In other words, the more robust scheme with low throughput like QPSK is chosen for low measured SINR value; this will be the case when UEs are located relatively far from the enodeb, in their region. On the other hand, a less robust technique with high throughput like 64-QAM is suitable for subscribers with high measured SINR values. Specifically, the selection mechanism depends on what is called the Channel Quality Indicator (CQI) which is reported by the UEs to the enodebs as one of the channel feedback reports. In LTE the CQI permits UEs to decode the data with a probability of error rate not to exceed 10%. The CQI parameter takes 15 possible values as shown in Table 3[21]. Both enodebs and UEs measure the received signal quality but only the enodebs make the selection decision for the modulation and coding schemes for both the uplink and the downlink, and higher efficiency will be achieved from higher modulation types. 25

36 Table 3. Channel Quality Indicator [21] 3.3 Multiple Access Techniques The previous schemes described above are used for one-to-one communications. In the real world, however, any base station must transmit signals to many UEs at the same time. Therefore, a technique called multiple access is used by the enodeb to share the resources of the air interface. 26

37 Different multiple access techniques are used by mobile communication devices; among them are the Frequency Division Multiple Access (FDMA) and the Time Division Multiple Access TDMA which are shown in Figure 11 [3]. In FDMA that was used by the first generation analogue systems, analogue filters were used by the UEs to differentiate their own carrier frequencies from other carriers. In this technique guard bands were used to separate adjacent carriers to minimize the interference between the two. On the other hand, in TDMA all mobile devices receive information on the same carrier frequency but at different times. Figure 11. Multiple Access Techniques [2] In the second generation cellular networks or GSM, a mix of FDMA and TDMA schemes are used for which every cell has a number of carrier frequencies; each one is shared by eight different mobiles. In the third generation another technique called Code Division Multiple Access (CDMA) is used in which the mobiles receive the information from the base stations on the same carrier frequency and at the same time but with unique codes. The mobiles use this code to distinguish their own original data from others. The LTE, on the other hand, uses a new technique which is a mixed technique from FDMA and TDMA called Orthogonal Frequency Division Multiple Access (OFDMA) for the downlink and Single carrier Frequency Division Multiple Access (SC-FDMA) for the uplink. Although LTE does not implement CDMA technique, it uses few of its concept for the control signals. 27

38 The difference between multiplexing and multiple access is that in the multiplexing system, the allocation of the resources is fixed while in the multiple access technique, the allocation of resources to mobile devices can dynamically be changed by the system. In other words, multiple access techniques are a more generalized form of the simpler multiplexing techniques TDD and FDD Modes In TDD both the base station and the UE transmit information on the same carrier frequency but at slightly different times while in FDD they both transmit at the same time, but they use different carrier frequencies. However, both techniques have pros and cons, In FDD usually the bandwidth is fixed and also the same for both uplink and downlink which makes it reliable for voice communications because the uplink and downlink data rates are alike. On the other hand, TDD is good for other applications in which the uplink and downlink data rates are different, like web browsing for which the downlink data rate is usually much higher than the uplink data rate. In this scheme the system decides how much time should be allocated for both uplink and downlink and adjusts its transmission time accordingly. Interference has a great deal of effect on TDD if one base station is receiving while a nearby station is transmitting. To avoid collisions and data loss, a careful time synchronization must be achieved so that all nearby base stations use the same time allocations for both uplink and downlink. In other words, all nearby base stations should transmit and receive at the same time. This scheme is reliable for isolated hotspot networks because each hotspot may have different time allocations. FDD, however, is preferred for wide-area networks with no isolation regions 28

39 3.3.2 Operation of TDD and FDD Modes In FDD full duplex mode operation, all mobiles must be equipped with a high attenuation duplex filter to isolate the uplink transmitter from the downlink receiver. In a half-duplex FDD mode, however, the base station can send and receive information at the same time but the UEs can do only one at a time. This means the UEs does not have to isolate the transmitter and the receiver which will facilitate the design of their radio hardware. In LTE all the modes mentioned above are supported. The base station can use either TDD or FDD modes and the UEs can support any combination of half duplex FDD, TDD or full-duplex FDD, but they can only use one technique at a time. Figure 12 [3] shows the operation of FDD and TDD modes. Figure 12. FDD and TDD Operation Modes [3] 29

40 3.4 Orthogonal Frequency Division Multiplexing (OFDM) In a multipath environment, high transmission data rates lead to a common problem called the Inter Symbol Interferences ISI [3]. OFDM is used for solving this issue by dividing the transmission of the information into parallel sub-streams by the OFDM transmitter instead of sending all the data as one stream. Each sub-stream is sent over a different frequency carrier known as the sub-carrier. Now the data rate on every subcarrier is less than before if the total data rate remains the same, and the symbol duration is now longer, so the ISI will be less than before and thus will reduce the error rate. Figure 13. Inter Symbol Interferences Reduction by Usage of Sub-Carriers [3] 30

41 An illustration of the ISI problem is shown in Figure 13[3]. The overall data stream of 400 Kilobits per second (kbps) is divided into four streams of 100 kbps; each one is sent over a unique sub-carrier (f1 to f4 in this example). The symbol overlap percentage is now only 10% if the delay spread between rays remains at 1 μs; thus the amount of ISI is reduced by one quarter of what it was before and the error rate at the receiver is reduced. In LTE, however, in OFDM the amount of sub-carriers can be as high as 128 for a 1.4 MHz frequency band and 2048 for the 20 MHz frequency band. Therefore, the ISI is reduced to negligible levels and it is no longer a problem Orthogonal Frequency Division Multiple Access Orthogonal Frequency Division Multiple Access technique is used in LTE in order to dynamically share the base station resources among the UEs as illustrated in Figure 14 [3]. In OFDMA, transmission happens at different times and different frequencies depending on the applications requirement. By looking at Figure 14, the type of application user1 is using is a voice over IP (voip), and hence these is no need for a high data rate, and thus number of carriers, but the flow of data should be constant. On the other hand, user2 is using non real time type of application that requires higher average data rate. However, the data comes in burst and the sub-carrier numbers can vary. User 3 is also a using a VOIP stream, but it is also affected by frequency dependent fading [3]. In OFDMA the base station deals with this issue by allocating the subcarriers that the UE is currently receiving a strong signal on and as the fading pattern changes, it will change the current allocation accordingly. It can also transmit using two blocks of sub-carriers separated by a fade as in the case for user 4. This dynamic allocation process that happens with any change in the fading patterns greatly reduces the damage of frequency and time dependent fading, and it requires UEs feedback. 31

42 Figure 14. Frequency and Time Implementation in OFDMA [3] The OFDMA access mode is used only with LTE downlink because it has a disadvantage that makes it unsuitable for the uplink: the high Peak to Average Power Ratio (PAPR). The transmitted signal power is subject to large variations. Figure 15 a [3] shows an example of a set of QPSK modulated sub-carriers with constant powers. Figure 15 b [3] shows a wide variations in the resulted amplitude signal with zero value when the peaks of the sub-carriers cancel and maximum value when they coincide together. These variations reflect the transmitted signals power as well as the transmission power amplifier as shown in Figure 15 c [2]. If the amplifier is linear then there will be a proportionality between the input and output waveform and the shape of the signal is not distorted. However, if the amplifier is not linear, then there will be no proportionality between the input and output signals and the resulting waveform will be distorted. This distortion in time domain will also cause distortion in the frequency domain, and the power of the signal will start leaking into adjacent frequency bands resulting in interference into other signals. 32

43 Figure 15. OFDMA waveform example. (a) Individual sub-carriers amplitudes. (b) Resulting OFDMA waveform amplitude. (c) Resulting OFDMA waveform power The principle of OFDMA work is the usage of Fast Fourier Transform (FFT) and Inverse Fast Fourier Transform (IFFT) in order to switch between the time domain representation of the signal and the frequency domain. Narrow orthogonal sub-carriers are used in the LTE OFDMA transmitter and each of them is 15 khz wide. The IFFT is used in the OFDMA transmitter for signals generation. Figure 16 [20] shows the OFDMA transmitter and Receiver and how the data travels from the modulator to the serial to parallel converter and from that to the IFFT unit. Every input to the IFFT resembles a specific sub-carrier. For the purpose of avoiding the inter-symbol interference problem, an additional block called a cyclic prefix follows the IFFT block. 33

44 Figure 16. OFDMA Transmitter and Receiver [20] What derives the need for OFDMA in LTE is the low-complexity design of the baseband receiver, compatibility with a variety of different bandwidths, usage of multiple antennas and a stable performance in the frequency selective fading channels Single Carrier Frequency Division Multiple Access The High PAPR problem in OFDMA makes it unreliable for LTE uplink because there is a theoretical possibility that all sub-carriers use the maximum power level and the combined powers of all sub-carriers will result in a very high power carrier. This can be dealt with at the downlink by using powerful expensive power amplifiers which are very close to linear at the base station. However, the UEs transmitters have to be cheap and power efficient to increase battery life, so this option is not available for LTE uplink. In order to reduce the PAPR and increase the efficiency of the power amplifiers and save 34

45 battery life, LTE uses a different access mode for the uplink called Signal Carrier FDMA (SC-FDMA). The principal form of SC-FDMA is similar to QAM modulation where the symbols are sent one at a time as it used to be done in Time Division Multiple Access (TDMA) systems in GSM. Figure 17 illustrates the block diagram of SC-FDMA transmitter and receiver [20]. Again for the purpose of eliminating inter-symbol interferences, a cyclic prefix block is being added as well. Figure 17. SC-FDMA transmitter and receiver with frequency domain signal generation [20] Figure 18 [22] illustrates the operational differences between OFDMA and SC- FDMA. The left side of the figure represents the OFDMA which is used in the downlink, and as discussed above, multiple-subcarriers are used and each of them is colored with a different color and modulated by a different data symbol, and each symbol lasts for a relatively long symbol duration. On the right side, the SC-FDMA represents the uplink. Although the signal carrier is actually a multiple carrier, all sub-carriers in the uplink are 35

46 modulated with the same data. It is shown in the figure by the fact that first green group of sub-carriers have the same color all around in the physical resource block, but obviously, it lasts for a much shorter time. This means that effectively the symbols duration are now much shorter and can be readily handled by regular cheap power amplifiers in UEs hardware. Figure 18. OFDMA VS SC-FDMA Access Modes [22] 3.5 LTE Air Interface protocol Stack Figure 19 [3] shows the used protocols in the LTE air interface from the UE viewpoint. It shows the flow of information between the protocol stack s different levels. Data packets are being created by the application in the user plane and processed by Transmission Control Protocol (TCP), User Datagram Protocol (UDP), or Internet 36

47 Protocol (IP). In the control plane, the signaling messages between the UE and the enodeb are exchanged, and these signals are written by the Radio Resource Control (RRC) protocol. The information from both cases is then processed by three protocols before being sent to the physical layer to be transmitted. These protocols are the Packet Data Convergence Protocol (PDCP), Radio Link Control (RLC) protocol, and Medium Access Control (MAC). Three parts form the physical layer: the transport channel processing unit is responsible for the error management procedure [3].The physical channel processing unit Figure 19. LTE Air Interface Protocol Stack [3] 37

48 is responsible for applying OFDMA, SC-FDMA techniques, and other functions [3]. The last part of the physical layer is the analog processer which is in charge of the conversion of the information to the analog form, filtering and mixing them up to the radio frequency for transmission [3]. The channels and signals represent the information flow between the different protocols in LTE. 3.6 LTE Channel Structure Three general channel types are used in LTE which are the logical, transport, and physical channels, and these channels are used to carry information between the different system layers in LTE. The channels are identified by the information they convey. Figure 19 shows the location of these channels in the LTE air interface protocol stack Logical Channels Data packets are transported through logical channels when traveling to the RLC layer from the MAC layer, as shown in Figure 19. Depending on the offered services, logical channels fall under two types [18]: traffic channels and logical control channels. A brief discussion of these channels follows below: Broadcast Control Channel (BCCH) It is in charge of sending system information to all the mobiles in the cell. This information is important for all UEs to manage themselves within the cell and to figure out the setup of the system [18]. Paging Control Channel (PCCH) This channel is responsible for paging any mobile whose location is unidentified in the network and the paging message has to reach more than one cell [18]. Common Control Channel (CCCH) It is in charge of the unified transmission of control information and random access [18]. 38

49 Dedicated Control Channel (DCCH) It sends control information to/from a mobile terminal, and it is also used for individual mobile configuration, such as handover messages [18]. Multicast Control Channel (MCCH) The control information required for the reception of the Multicast Traffic Channel (MTCH) is being sent by this channel [18]. Dedicated Traffic Channel (DTCH) This logical channel carries all the data to/from the mobiles. All non-mobile (Multicast/Broadcast Signal Frequency Network) (non-mbsfn) data in the downlink and all the uplink user data are transmitted in this channel [18]. Multicast Traffic Channel (MTCH) This channel in in charge of the downlink transmission of the Multimedia Broadcast Multicast Service (MTCH) Transport Channels The transport channels provide services from the physical layer to the MAC layer. This channel is defined by the How and with what characteristics the data should be transmitted over the radio interface [18]. Below is a brief discussion of these channels: Broadcast Channel (BCH) Parts of the BCCH system information are being sent by this channel over the network coverage area [18]. Paging Channel (PCH) Paging information from the PCCH logical channel is sent by this channel. This channel also saves UEs battery life by the Discontinuous Reception (DRX) and by waking up at only predefined times [18]. 39

50 Downlink Shared Channel (DL-SCH) This channel is the main one responsible for LTE downlink data transmission, including both control and traffic data. It is compatible with multicast/broadcast transmission, Hybrid Automatic Repeat Request (H-ARQ), dynamic rate adaptation, spatial multiplexing, and DRX for reducing UEs power consumption [18]. Multicast Channel (MCH) This channel supports Multicast/Broadcast Single Frequency Network (MBSFN) transmission and is used in association with MCCH and MTCH logical channels. The MBSFN sends the information on the same frequency to multiple terminals which are connected to many synchronized enodebs [18]. Uplink Shared Channel (UL-SCH) This channel job conveys uplink data in LTE, just as DL-SCH does in the downlink [18]. Random-Access Channel (RACH) Although this channel does not convey any transport blocks, it is still considered as a transport channel and its responsibility is to initiate access to the network [18] Physical Channels These channels are in charge of the real transmission of data that occurs between the mobile and the cellular tower or enodeb in the LTE system [18]. Resource blocks and resource elements are the fundamental components of the physical channels. There are two groups of physical channels in the LTE system in terms of mapping with transport channels. One of these groups can relate with the transport channels, but this is not the 40

51 case with the other group which is known as the L1/L2 control channels which exist on both the uplink and the downlink. They provide the information for receiving and decoding the signal properly on the downlink, and this information is called Downlink Control Information (DCI). On the other hand, they provide Uplink Control Information for the H-ARQ protocol and the scheduler on the uplink [18]. Below is a brief description of the physical channels used in the LTE system: Physical Downlink Shared Channel (PDSCH) The task of this channel is the same as the task of the DL-SCH and PCH transport channels, as it is in charge of sending UEs data on the downlink [18]. Physical Broadcast Channel (PBCH) Its responsibility is to carry partial system data for allowing the user accessibility to the network [18]. Physical Downlink Control Channel (PDCCH) It is mainly in charge of transmitting scheduler decisions, which are part of the downlink control information [18]. Physical Multicast Channel (PMCH) It is mainly in charge of MBSFN transmission [18]. Enhanced Physical Downlink Control Channel (EPDCCH) It has more flexibility in allowing control information, but it has the same job as PDCCH and it was defined in Release 11 [18]. Physical Hybrid-ARQ Indicator Channel (PHICH) This channel is responsible for sending H-ARQ acknowledgment to the mobile terminal as an indication of whether the data should be resent by the enodeb or not [18]. Relay Physical Downlink Control Channel (R-PDCCH) This channel that was defined in Release 10, has the responsibility to carry out L1/L2 control information from the base station to the relay link [18]. Physical Uplink Shared Channel (PUSCH) As the name implies, this channel s task is the opposite of the PDSCH in the uplink; it sends the mobile uplink data to the base station [18]. Physical Control Format Indicator Channel (PCFICH) This channel supplies the information required for PDCCH decoding to the mobile terminal [18]. 41

52 Physical Random Access Channel (PRACH) It is in charge of sending the random access requests the mobile terminals make [21]. Physical Uplink Control Channel (PUCCH) H-ARQ acknowledgments, Channel Quality Indicator (CQI), uplink scheduling requests and others are carried out by this channel which is in charge of conveying the uplink control information [18]. Some of the physical channels like (PDCCH, R-PDCCH, PCFICH, EPDCCH, and PHICH) that are used for the downlink control information do not have any mapping to the transport channels [18]. 3.7 LTE Measurements Any cellular Drive Test (DT) tool used for LTE spectrum evaluation reports many different measurements, among which are very important ones such as Reference Signal Received Power (RSRP), Received Signal Strength Indictor (RSSI), Reference Signal Received Quality (RSRQ), and Signal to Interferences plus Noise Ratio (SINR). Many critical decisions depend on these measurement values such as in handover situations. A brief description of each one is illustrated below: Reference Signal Received Power (RSRP) The mean power of the Resource Elements (RE) which convey cell Reference Signals (RS) over the whole bandwidth is defined as RSRP. RSRP values range from -140 dbm to 44dBM with resolution of 1 db. Mapping of reported RSRP values with measured quantity are shown in Table 4 [23]. Handover process, cell selection and resection in the network relies heavily on the RSRP reported readings to the UE, which keeps measuring RSRP from the serving cell and from the neighboring cells as well. RSRP provides information about the signal power but not the signal quality. 42

53 Table 4. RSRP Measurement Reports Mapping [23] Received Signal Strength Indicator (RSSI) The overall power, including the power of the serving cell, interference, and noise power received by the UE over the whole LTE channel is defined as RSSI. RSSI helps in determining noise and interference information. Therefore, RSRQ measurement depend on both RSRP and RSRQ [23]. Reference Signal Received Quality (RSRQ) The signal quality is determined by this parameter, which ranges from db to -3 db with a half db step. The mapping of reported RSRQ values is shown in table 5 [23]. RSRQ can be calculated using the equation shown in (1). It is obvious from the equation that RSRQ depends on RSRP, RSSI, and the number of resource blocks [23]. RSRQ= (RSRP/RSSI) NRB (1) 43

54 Table 5. RSRQ Measurement Reports Mapping [23] Signal to Interference plus Noise Ratio (SINR) SINR is calculated using the equation shown in (2) [23]. S represents is the signal power, I is the average interference and N is the Noise power. All are measured over the same bandwidth. SINR= [S / (I+N)] (2) 44

55 Chapter Four Measurement Procedure 45

56 4.1 Introduction The performance of cellular networks can be determined through a common method known as the drive-test. Drive tests measure received signals from fixed beacon channels, such as Broadcast Common Control Channel (BCCH) in Global System for Mobile communication (GSM, second generation, 2G), or Common Pilot Channel (CPICH) in Wideband Code Division Multiple Access (WCDMA, third generation, 3G). In Long- Term Evolution (LTE) networks of fourth generation (4G), cell beacons are referred to as Reference Signals. Tools used in drive tests can be classified into three categories: phonebased devices, scanners, and advanced receivers with high dynamic range. A similar study conducted on a GSM network used the phone-based tool and has shown low repeatability [24]. Since the phone needs to be inside a vehicle, the phone-based devices do not provide accurate measurements, due to the in-vehicle penetration losses as reported in [25]. The same study showed that both the scanning receiver and the high dynamic range receiver exhibit similar behavior, albeit with different levels of reliability and repeatability. In this study, the scanning receiver is used to measure the RSRP in a commercial network. Many factors affect the propagation of radio-frequency electromagnetic waves in terrestrial channels, such as atmospheric conditions, multipath fading environment, scattering, and others [26]. All these effects alter the wireless channel properties and influence the repeatability of drive-test outcomes. Therefore, different measurements can be expected even in this same location because of the unpredictable varying propagation conditions and fading. The goal behind this study is to estimate the impact of these fluctuating factors on the repeatability of LTE RSRP measurements. This will help to establish limits on accuracy and standard deviations that can be expected in LTE propagation modeling. Two important concerns need to be verified: first, whether or not the RSRP values are repeatable, or, in other words, if there is a bias between different drive-test measurements taken in the same location on different days; second, to quantify bias variability, if the biases between different drives vary. 46

57 This chapter is organized as follows. Following this introductory section, Sections 4.2 and 4.3 explain the measurement equipment and the drive-test setup. Collected data analysis is presented in Sections 4.3, 4.4, and 4.5, and the chapter that follows summarizes the conclusion and suggests future work. 4.2 Measurement Tools Measurement setup used to collect data for this study includes common drive-testing equipment widely used by RF engineers in the cellular communications industry. The equipment list includes a RF scanner, laptop with scanner software, USB cable, GPS receiver, external RF antennas, 12V-5V DC power converter, and a vehicle that will carry all the above equipment during the test. The main unit used in this system is the multi-technology (2G/3G/4G) SeeGull EX+ scanning receiver, kindly shared by the PCTEL Inc. This tool is used for modulation and signal strength measurement, engineered and designed for the rigors of cellular networks inspection during planning, assembly and maintenance of wireless networks. Scanner application scenarios include acceptance testing, competitive benchmarking, spectrum clearing, troubleshooting, and network optimization. The front view of this scanner illustrating all connections is shown in Fig.1 [27]. Advanced signal processing techniques are used in SeeGull scanners. The Carrier to Interference plus Noise Ratio (CINR) dynamic range for this type ranges from -20 to + 40 db. Valid RSRP measurements as low as -140dBm minimum detection level can be reported [28]. This high dynamic range allows collecting more measurement points compared to the phone-based system. The SeeGull scanner supports all standard LTE bandwidths (1.4, 3, 5, 10, 15 and 20MHz) and multiple data modes up to 6GHz. Depending on the multiple-input-multiple-output (MIMO) mode, up to 48 reference signals can be measured per second. In addition to GSM/WCDMA/LTE scanning mode, SeeGull can be used in the spectrum analyzer mode. Spectrum analysis can detect center frequency of LTE carriers. 47

58 Figure 20. Front View of SeeGull EX+ Receiver [27] 48

59 As seen in Figure 20, the external RF antennas are connected to the scanner through three SMA connecters and placed on top of the vehicle, removing the impact of invehicle penetration losses. One GPS receiver, one USB cable connected to a laptop with appropriate measuring software, and 12V to 5V DC power cable are all connected to this unit as well. GPS position accuracy is estimated at ±2.5m. The Windows-based laptop is used to log the scanner measurements. 4.3 Drive Test Procedure The drive-tested area for this study was a section of Melbourne, Florida, the home city of Florida Institute of Technology. The measurements were taken on five main roads and two branch streets over three consecutive days. The total route length was approximately 25 miles. This cluster can be classified as the suburban environment. Terrain is flat and the majority of structures in this area have either one or two stories. All RSRP measurement points had been recorded on the laptop s hard-drive, along with their GPS coordinates by the scanner software, as the test-vehicle passed throughout the grid. The cellular coverage of the section under the test was provided by 24 cells from a leading nationwide LTE operator. Fig. 21 shows the GPS-logged path on a street map. 49

60 Figure 21. Route Map 50

61 4.4 Measurements Analysis The SeeGull scanner conveniently exports individual drives integrating GPS information together with RSRP and physical cell ID (PCI) into text files. PCI in LTE is a combination of 3 primary and 168 secondary synchronization codes (504 PCIs total) and identifies the strongest and all other cells detected at a GPS given location. All logged georeferenced RSRP measurements were exported from the scanner and imported into MS Excel and MATLAB for a statistical analysis. The aim of this analysis has been to evaluate repeatability of RSRP values from two aspects. First, deciding if the obtained measurements are in general repeatable on a dayby-day basis. General repeatability is tested by comparing the averaged geographically binned RSRP biases between different daily drives. Second, calculating the measurement scatter and the probability of RSRP differences falling within a particular interval. Three main software packages are used: 1. SeeHawk DT software: comes with the SEEGUL EX+ scanning receiver from PCTEL Inc for recording all measurements during the test on the laptop s hard drive. 2. Microsoft Excel : The logged data from the drive test is fed into Excel for data binning (which will be explained in the following section) as well as statistical and mathematical operations. 3. MATLAB : The processed information from Excel is imported into MATLAB, which is a high technical language used for further data computations, analysis, and final representation. This is done using strings of codes developed in MATLAB. 51

62 4.4.1 Data Binning To facilitate the statistical analysis, the drive-tested region has been divided into a fixed grid with 50 squared-meter bins to average-out the fast-fading effect of instantaneous RSRP measurements. After gridding data into a sparse matrix resembling the drive route on Figure 21, local means of all RSRP measurements are calculated using the logarithmic averaging within each bin shown in (3). Each bin is a unit geographic area with a square shape. Figure 22 [24] illustrates the binning idea. The RSRP measurements are sampled by the scanner from the surrounding cells in the area within each bin. These samples are depicted in the figure as red dots along the drive route. Ni RSRPi = 1 RSRP Ni k=1 k (3) RSRPi is the averaged reference signal received power in a specific bin; RSRPk is the measured signal strength collected in the i-th bin and it is expressed in dbm; Ni represent the overall number of RSRPk in a specific i-th bin. The location of each bin is identified by the longitude and latitude coordinates that define the bin geographical position and is assumed to be at the center of the bin. The number of measurements per bin has been used to additionally filter bins with less than 5 RSRP readings. Filtering out bins with low measurement counts is expected to reduce the measurement noise by eliminating bins with few data points that are a likely source of increased bias variance. 52

63 Figure 22. Binning Idea [24] Averaged RSRP Analysis At least three days of drive testing must be conducted to be able to determine if there is a repeatability in the RSRP measurements or not. In other words, this is done to find out if the average RSRP measurements for the different drive tests are maintained throughout the given route. Obviously, it is normal that any two geographical positions with the same latitude and longitude coordinates can have different instantaneous RSRP readings due to many reasons, among which are the atmospheric conditions and their effect on the propagation of electromagnetic waves in terrestrial channels. However, if these instantaneous measurements are averaged in each bin and only the averaged values are compared, then the general trend of the averaged RSRP values throughout the route should be maintained, and this is how repeatability of RSRP measurements is tested in LTE. This will only be true if no changes were made to the base station transmitting 53

64 power or other parameters or factors that can alter the propagation of the cellular LTE signals, including atmospheric conditions. The comparison of averaged RSRP readings between different daily drives is done on a bin-by-bin basis. After gathering all the data from all three days, it has been determined that RSRP measurements originate from a total of 24 sectors that cover the driven region on LTE channel 5780 [29]. Therefore, a two-dimensional space with 3 X 24 (=72) data sets has been created, each data set representing measured data from one day and one cell. Next, the mean of all RSRP values from each cell on each day and in each bin is calculated using logarithmic averaging (in decibel domain). Only bins with measurements on multiple days are considered to enable bias comparisons and calculations. The entire process is repeated again, but now only bins that have at least 5 RSRP measurements within are considered. The reason behind this is to remove any bins with few RSRP measurement points to help eliminates the outliers that might affect the accuracy of the result. Another reason behind the last step is to compare the two process s outcome and note if there is a significant variance in data analysis when eliminating bins with fewer measurements. Log-averaged RSRP values from an arbitrary subset of 100 bins, are presented in Figure 23 (all bins) and Figure 24 (excluding bins with less than 5 measurements), respectively. 54

65 Figure 23. Averaged RSRP for a Subset of Geographical Bins (All Bins) 55

66 Figure 24. Averaged RSRP for a Subset of Geographical Bins (Only Bins with More than 5 Readings) As expected, the RSRP averages in the second plot fluctuate less after filtering lowcount bins. Minor daily fluctuations are also observed. Signal levels recorded in Figure 23 and Figure 24 are relatively high and reflect the outdoor nature of the drive test, relatively flat morphology, and the 700MHz LTE carrier that propagates further than upper band carriers. These levels do not coincide with a typical RSRP actual traffic distribution. Actual LTE traffic is expected to be carried much more indoors with reduced RSRP levels, mostly around and below -100dBm. However, 56

67 increased RSRP levels are not critical in order to evaluate repeatability. Biases between measurements taken during different daily drives do not depend on the signal level Comparison of RSRP Measurements It is easier to understand how the comparison is done by imagining each bin having three averaged RSRP readings with every value being bonded to one of the days. Thus, to determine repeatability the pairwise differences are calculated. In other words, three biases between RSRP measurements from different days are evaluated for each bin using the equation shown in (4). ɨɉ = RSRPɨ - RSRPɉ, ɨ = 1, 2, 3: ɉ = 1, 2, 3: ɨ ɉ (4) In equation (4), RSRP denotes the averaged value (local mean) calculated through logarithmic averaging on a given day in the observed geographical bin. A matrix of differences is formed from (4) for each cell as shown in equation (5). Each in equation (5) is a matrix with a two-dimensional RSRP grid, where superscript denotes cell index (from 1 to N=24) and subscripts denote days for bias calculation. Thus, equation (5) represents a 3-by-24 matrix of georeferenced matrices containing RSRP biases between pairs of daily drives: (5) N 12 N 13 N 23 The differences of averaged RSRP values in bins between the first day and the second day are shown in column 1, Column 2 represents the differences between day 1 and day 3, and finally column 3 represents the averaged RSRP differences in bins 57

68 between the second and third days. In other words, each row in the array shows the differences of averaged RSRP values between different days for all bins. Examples of pair-wise biases between different drives considering all bins and bins with at least 5 readings associated with the measurements in Figure 23 and Figure 24 are plotted in Figure 25 and Figure 26 respectively. These plots show clear centering of biases around 0dB with reduced spread in Figure 26 after filtering bins with low measurement counts. Some isolated bins show biases as high as 5dB but overall spread points to apparent repeatability of RSRP measurements in the tested morphology. The matrix elements described in equations (4) and (5) are considered as random variables. Therefore, to determine if there is a significant bias between different drives, it is necessary to test the central tendency of all the differences. This is accomplished by taking averages and standard deviations of biases between different days. Means grouped around zero and small standard deviation not exceeding 1 db imply that the RSRP measurements are repeatable on average. Otherwise, there is a bias between different days with significant implications on the propagation modeling and optimization processes that rely mostly on single measurement campaigns. 58

69 Figure 25. Pair-wise RSRP Differences for a Subset of All Bins 59

70 Figure 26. Pair-wise RSRP Differences for a Subset of Bins with at Least 5 RSRP Measurements 60

71 Another way of the repeatability verification is the test known as the ANalysis Of VAriance (ANOVA) method. It has been applied to the three bias data sets in (4). Grouping of means near zero db indicates that the averaged RSRP differences in each bin is unbiased for all drives and the scanner drive-test device maintains average repeatability. In statistics, this is called satisfying the null hypothesis (H0) condition which indicates that all the means of the columns in array (5) are coming from the same population. In order for the ANOVA test outcome to be accurate, three assumptions are assumed: All observations (averaged RSRP biases in this case) are mutually independent. All observations have the same variance. All observations are normally distributed Results for Comparison of RSRP Measurements ANOVA1 graphs for the scanner DT tool taking into account all bins and bins with more than 4 readings are presented in Figure (27) and Figure (28) respectively. Each figure has three boxes and whisker plots. The plots are associated with the pair wise differences from (4). In other words, each plot represent one of the columns in (5). The first plot from the left specify the differences between measurements collected on the first and the second day, while the second one represents the biases between the first and the third and finally the last one exemplifies the differences between the second and the third day. Median of all biases is close to 0dB in all daily drive pairs on both graphs, indicating centering biases around 0dB as expected. The black-colored whiskers at approximately ±4dB, represent the maximum and minimum values of all RSRP differences which are not considered as outliers. The end of the blue flat lines represents the interquartile range. The upper blue flat lines represent the 75% percentile of the data and the lower blue flat lines represent the 25% percentile of the data. This means 50% of measurements are contained with the rectangular boxes (approximately ±1dB from 0dB). All red + marks are representing outliers that can be discarded from the statistics. 61

72 Figure 27. ANOVA Graph for a Subset of All Bins Table 6. ANOVA Table for a Subset of All Bins 62

73 Figure 28. ANOVA Graph for a Subset of Bins with at Least 5 RSRP Measurements Table 7. ANOVA Table for a Subset of Bins with at Least 5 RSRP Measurements 63

74 A significant outcome of the ANOVA1 test is called the p-value that indicates if there is a rejection of the null hypothesis H0 condition. In other words, it will determine if there is a difference between the means of each plot observations. If the p-value is far less than 0.05 (an alpha of 5%), then this implies that at least one of the means from one plot is different from the others. In addition to the p-value and the graph, the ANOVA1 returns a table and both the graph and the table are difficult to interpret at first glance. Therefore, it is necessary to explain the two. The variability in the differences of the observations can be divided into two categories for each graph: Variability between groups: is the variability among the means of all the values represented by the three plots that correspond to the three columns in (5). Variability within groups: is the variability among the differences within each plot that correspond to the particular column in the matrix shown in (5). In addition to the graph, the ANOVA function returns a table which has six columns: 1. The first one represents the source of variability. 2. The second column has three values: sum of squares between groups, Sum of Squares Within groups (SSW) and Total Sum of Squares (SST). It is easier to start with calculating the sum of squares within groups by first calculating the mean of all RSRP measurements for the first sample then calculating the difference between each observation and the mean and square the outcomes and add them all up as shown in (6). This step is again repeated for the other two different samples. The sum of squares within groups referred as Error in the ANOVA table is equal to the sum of squares of each of the individual groups. ( Observation mean )^2 (6) The Total Sum of Squares (SST) is calculated by taking all the three data samples treating them as one sample. The next step is calculating the mean value of the resulted sample, then calculating and squaring the variance between each individual value and the mean and then summing all the outcomes together to get 64

75 the total sum of squares (SST), referred as Total in the ANOVA1 table. Finally, the sum of squares between groups is calculated algebraically by subtracting the SSW value from SST value. Another way to calculate the sum of squares between groups is by subtracting the mean from the sample, where all three samples are combined, from the mean of each individual sample, square each of these values and add them all together and finally multiply the outcome by the number of measurements in any individual sample to get the sum of squares between groups. 3. The third column is the degree of freedom (df) for each source. The degree of freedom for the sum of squares between groups can be calculated using the equation shown in (7) and the degree of freedom for the SSW is evaluated using the equation in (8): df ( sum of squares between groups) = no. of groups -1 (7) df ( sum of squares within groups) = observations no of groups (8) 4. The fourth column is the Mean Squares (MS) associated with each source. The MS for the sum of squares between groups can be calculated using the equation in (9) and the MS for SSW is evaluated using (10) as follows: MS ( sum of squares between groups ) = (9) MS ( sum of squares within groups ) = (10) sum of squares between groups degrees of freedom in (7) sum of squares within groups degrees of freedom in (8) 5. The fifth column represent the F statistic that can be calculated using the ratios of the MS s and using the equation shown in (11): F statistic = MS ( sum of squares between groups) in (9) MS ( sum of squares within groups ) in (10) (11) 65

76 6. The last column represents the p-value that is derived from the cumulative distribution function (cdf) of F. As F value goes up p-value goes down. From the degree of freedom shown in (7), the degree of freedom shown in (8), the level of significance ( ) and the value of F one can look up the critical value for an of 5% using the F distribution table shown in (8) [30]. The final step after finding the critical value is drawing the distribution marking the critical value with a red line as shown in figure 29. Every value to the left of the critical value is in the green area and everything to the right, or greater than the critical value, is in the rejection area. Now it is clear to know which region the resulted F value may fall in. The critical value in this study for the first case scenario (taking all bins into account) according to table (8) is 3 and the F value is The critical value in the second scenario, when considering only bins with more than four measurements, is also 3 but the F value is It is obvious that both F values are greater than the critical value which means they are both located on the right side of the distribution shown in Figure 29. Therefore, theoretically saying, in this study the null hypothesis (H0) is rejected for both cases. Significant differences among the center lines of the box-and- whiskers in the plots results in large F values that lead to small p-values. As seen from tables 6 and 7 the scanner express very small p-values (less than 0.05) implying in the rejection of H0, that there is a variance in the mean RSRP values of the differences among the three drives. However, these small variations (all RSRP biases are within 0.5 db from each other in this study) will not affect LTE scanner RSRP repeatability because the scanner precision is 1dB. 66

77 Table 8 67

78 Figure 29. F-distribution curve In a nut shell, ANOVA plots identify two main observations. Firstly, the differences between measurements gathered on different days have no significant biases. Secondly, besides removing some outliers, there is no significant bias reductions after eliminating bins with less than 5 readings from the statistic. This means that although a number of bins with few measurement points are excluded, their biases are not significant to affect overall statistics. This may be justified by the fact that the scanner already logs RSRP values as a weighted average of multiple successive instantaneous measurements. Therefore, additional measurement spread reduction by excluding bins with low measurement count does not appear necessary for the LTE scanner RSRP measurement set. Overall biases between drives are within 0.5dB pointing to the conclusion that LTE scanner RSRP measurements are in general repeatable, and the single drive test, without additional averaging of multiple drives, may be deployed towards RF optimizations and benchmarking 68